Data collection systems are used for material analysis and microanalysis of a variety of material properties including chemical, structural, mechanical, crystallographic, or other information. For example, Energy Dispersive Spectrometry (“EDS”) has grown into a robust analytic technique for the measurement of material properties. EDS is an analytical technique performed in a scanning electron microscope (“SEM”) or transmission electron microscope (“TEM”) in a low pressure or near vacuum environment. A sample is positioned beneath a column housing an electron source. The electron source may be any suitable source, such as a tungsten filament, thermal field emission, or LaB6 electron source. The electron source may emit electrons that are directed in a beam through the column and toward a sample chamber. The sample chamber may be connected to the column and allow a sample to be held in line with the electron beam for imaging and/or sampling. The sample may have an unprepared surface allowing sampling of the exposed surface (such as particles or broken and/or cut surfaces) or a prepared surface that is substantially flat. Non-conductive samples may be made more conductive by deposition of a conductive layer over at least part of the surface in order to provide a conductive path to ground. For example, carbon layers or gold layers sputtered onto the surface of a sample can provide a conductive layer that dissipates charge from the sample to the sample stage or other ground within the sample chamber.
Referring now to FIG. 1, conventional EDS may be conducted in an SEM 100 by presenting a sample 102 at in line with an electron beam 104. The surface of the sample 102 may be oriented perpendicularly to the electron beam 104 or may be oriented at an angle not perpendicular to the electron beam 104. For a sample 102 with an uneven surface, tilting of the sample provides line-of-sight to features that are otherwise inaccessible by the electron beam 104. The position of the sample 102 relative to the beam 104 may be achieved by tilted a sample stage 106 or by providing a sample holder (not shown) having non-parallel surfaces mounted to the sample stage 106 or a combination of the two.
Lenses, such as electromagnetic lenses, may focus and/or deflect the electron beam 104 at different working distances (focal length beneath a lowest point of the column 108) and/or locations on the sample 102. A “scan” of the SEM 100 may include construction of an image of a surface of the sample 102 by rastering the beam 104 through a predetermined range of positions and/or deflections of the beam 104. A combination of an EDS detector 110 and rastering of the beam 104 allow for the construction of X-ray count maps of a portion of the sample 102.
The interaction of the electron beam 104 and the sample 102 causes the atoms of the sample 102 to become excited. When an electron or electrons of an atom relaxes to a lower-energy ground state, the atom will emit energy in the form of an X-ray. The X-ray will have a particular energy that correlates to the state of the electron that emitted the X-ray. For example, electrons in the K energy level of the atom will emit an X-ray with a different energy than electrons in the L energy level. The X-rays will also vary in energy depending on the element emitting the X-ray. For example, electrons of the K energy level in aluminum will emit X-rays of different energy than the electrons of the K energy level in iron. Measurement of the X-ray energy allows for differentiation of elements excited by the electron beam 104. The relative quantity of X-ray counts in a given period of time indicates relative concentration of those elements in the sample 102 excited by the electron beam 104.
The EDS detector 110 includes a detection surface that converts X-rays into a voltage signal. The voltage signal is the provided to a pulse processor that measures the signal and passes them to an analyzer, which will then display the data and allow further analysis by a user. The detection surface can be a semiconductor that is cooled to low temperatures, for example, by liquid nitrogen or by Peltier cooling. EDS detectors include silicon-lithium (“Si(Li)”) detectors and newer silicon drift detectors (“SDDs”).
Rastering the electron beam 104 across the surface of a sample 102 allows for the collection of X-ray count maps of the surface. The X-ray maps can include individually selected energy channels or each data point within the map can include a full spectrum for the point. Calculating the relative concentrations of various elements in the sample 102 is performed by comparing the relative intensities of energy channels having local maximum in the X-ray spectrum of each point.
During X-ray mapping of a sample 102, the individual sampling locations of the surface have relatively low quantities of X-ray counts detected by the EDS detector 110. The relatively low X-ray counts result in low-resolution spectra and/or poor statistical quality of the elemental identification. The EDS detector 110 is limited by the physical collection area of the detection surface in the EDS detector 110 and by the energy resolution that the EDS detector 110 can maintain as throughput increases. Recent advancements in EDS detectors 110 have allowed increased collection rates of X-ray counts and SEM 100 improvements have allowed for greater current to be applied to the sample 102 by the electron beam 104. The increased collection rates of X-ray counts create a greater discrepancy in the statistics when the electron beam 104 falls onto a portion of the sample 102 generating lower count rates. The count rate for a particular sampling location is affected by the sample composition (i.e., atomic weight of the elements present), sample geometry, and beam conditions.
In conventional EDS mapping and other data collection techniques for material analysis, the dwell time per sampling location is set at the beginning of the mapping collection run, and the dwell time is constant for the duration of the mapping. The dwell time is set based on user-selected sampling locations during setup of the mapping; the dwell time is a user-specified estimate of the time needed to attain satisfactory statistics at each sampling location. While data collection rates have increased in recent years, improvements to statistics, particularly early in the scan of a sample 102, are desirable.